Data as of Jul 25, 2026 · Based on 2,718,867 AI responses across 9,511 prompts · See how Parse measures this
eRAGgithub.comeRAG is a novel evaluation approach for retrieval-augmented generation (RAG) systems that assesses each retrieved document's individual contribution to downstream task performance. It achieves higher correlation with actual RAG performance than traditional methods while using up to 50 times less GPU memory.
Parse Score
